Unverified50% confidenceFactExact time
A so-called "uncensored" model means it lacks safety alignment
1
Sources
50%
Confidence
Long-term
Relevance
9/15/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedModels processed through abliteration suffer almost no loss in general capability, yet lose most of their safety constraints.71% similarUnverifiedRelying on prompt constraints to prevent dangerous operations is inherently unreliable, because the model may be jailbroken, make misjudgments, or forget constraints during long conversations.71% similarUnverifiedModel-level safety alignment is fragile, because open-source model weights are accessible and behaviors are modifiable, meaning the alignment efforts invested by vendors can be erased by simple downstream operations.70% similarUnverifiedAgent评测框架处于无人认领、逐渐drift（漂移）的状态比没有框架更危险，因为它会给团队虚假的安全感68% similarUnverifiedSafety-aligned models are theoretically supposed to refuse high-risk requests involving weapons manufacturing, violent attacks, and similar topics.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/923506API
curl https://kongchang.com/api/v1/knowledge/claims/923506MCP
get_claim(id=923506)